diff --git a/apachesolr.api.php b/apachesolr.api.php index d81306b..57a50aa 100644 --- a/apachesolr.api.php +++ b/apachesolr.api.php @@ -194,17 +194,56 @@ function hook_apachesolr_delete_by_query_alter($query) { $query .= ' AND hash:' . apachesolr_site_hash(); } } +/* + * This is the place to look for the replacement to hook_apachesolr_node_exclude + * You should define a replacement for the status callback and return + * FALSE for entities which you do not want to appear in the index and TRUE for + * those that you want to include + */ + +/** + * This is invoked for each entity that is being inspected to be added to the + * index. if any module returns TRUE, the entity is skipped for indexing. + * + * @param integer $entity_id + * @param string $entity_type + * @param integer $row + * A complete set of data from the indexing table. + * @param string $env_id + * @return boolean + */ +function hook_apachesolr_exclude($entity_id, $entity_type, $row, $env_id) { + // Never index media entities to core_1 + if ($entity_type == 'media' && $env_id == 'core_1') { + return TRUE; + } + return FALSE; +} + +/** + * This is invoked for each entity from the type of ENTITY_TYPE that is being + * inspected to be added to the index. if any module returns TRUE,  + * the entity is skipped for indexing. + * + * @param integer $entity_id + * @param integer $row + * A complete set of data from the indexing table. + * @param string $env_id + * @return boolean + */ +function hook_apachesolr_ENTITY_TYPE_exclude($entity_id, $row, $env_id) { + // Never index ENTITY_TYPE to core_1 + if ($env_id == 'core_1') { + return TRUE; + } + return FALSE; +} /** * Add information to index other entities. * There are some modules in http://drupal.org that can give a good example of * custom entity indexing such as apachesolr_user_indexer, apachesolr_term * - * This is the place to look for the replacement to hook_apachesolr_node_exclude - * You should define a replacement for the status callback and return - * 0 for nodes which you do not want to appear in the index and 1 for nodes - * that you do. See http://drupal.org/node/1474906 for an example. - * * @param array $entity_info */ function hook_apachesolr_entity_info_alter(&$entity_info) { diff --git a/apachesolr.index.inc b/apachesolr.index.inc index ea8d2a2..cee8296 100644 --- a/apachesolr.index.inc +++ b/apachesolr.index.inc @@ -10,14 +10,32 @@ * Send up to $limit entities of each type into the index. */ function apachesolr_index_entities($env_id, $limit) { - $entities_processed = 0; + $documents_submitted = 0; foreach (entity_get_info() as $entity_type => $info) { // With each pass through the callback, retrieve the next group of nids. $rows = apachesolr_index_get_entities_to_index($env_id, $entity_type, $limit); $documents = array(); foreach ($rows as $row) { if (!empty($row->status)) { - $documents = array_merge($documents, apachesolr_index_entity_to_documents($row, $env_id)); + // Let any module exclude this entity from the index. + $build_document = TRUE; + foreach (module_implements('apachesolr_exclude') as $module) { + $exclude = module_invoke($module, 'apachesolr_exclude', $row->entity_id, $entity_type, $row, $env_id); + // If the hook returns TRUE we should exclude the entity + if (!empty($exclude)) { + $build_document = FALSE; + } + } + foreach (module_implements('apachesolr_' . $entity_type . '_exclude') as $module) { + $exclude = module_invoke($module, 'apachesolr_' . $entity_type . '_exclude', $row->entity_id, $row, $env_id); + // If the hook returns TRUE we should exclude the entity + if (!empty($exclude)) { + $build_document = FALSE; + } + } + if ($build_document) { + $documents = array_merge($documents, apachesolr_index_entity_to_documents($row, $env_id)); + } } else { // Delete the entity from our index if the status callback returned 0 @@ -27,7 +45,7 @@ function apachesolr_index_entities($env_id, $limit) { $indexed = apachesolr_index_send_to_solr($env_id, $documents); if ($indexed !== FALSE) { - $entities_processed += count($rows); + $documents_submitted += count($documents); $index_position = apachesolr_get_last_index_position($env_id, $entity_type); $max_changed = $index_position['last_changed']; $max_entity_id = $index_position['last_entity_id']; @@ -45,7 +63,7 @@ function apachesolr_index_entities($env_id, $limit) { apachesolr_set_last_index_updated($env_id, REQUEST_TIME); } } - return $entities_processed; + return $documents_submitted; } /** @@ -184,11 +202,6 @@ function apachesolr_index_entity_to_documents($item, $env_id) { * @return number indexed, or FALSE on failure. */ function apachesolr_index_send_to_solr($env_id, $documents) { - // Do not index when we do not have any documents to send - if (empty($documents)) { - return FALSE; - } - try { // Get the $solr object $solr = apachesolr_get_solr($env_id); @@ -201,7 +214,11 @@ function apachesolr_index_send_to_solr($env_id, $documents) { watchdog('Apache Solr', nl2br(check_plain($e->getMessage())), NULL, WATCHDOG_ERROR); return FALSE; } - + // Do not index when we do not have any documents to send + // Send TRUE because this is not an error + if (empty($documents)) { + return TRUE; + } // Send the document off to Solr. watchdog('Apache Solr', 'Adding @count documents.', array('@count' => count($documents))); try { @@ -339,8 +356,9 @@ function apachesolr_index_get_entities_to_index($env_id, $entity_type, $limit) { // Find the next batch of entities to index for this entity type. Note that // for ordering we're grabbing the oldest first and then ordering by ID so // that we get a definitive order. + // Also note that we fetch ALL fields from the indexer table $query = db_select($table, 'aie') - ->fields('aie', array('entity_type', 'entity_id', 'changed', 'status')) + ->fields('aie') ->condition('aie.bundle', $bundles) ->condition(db_or() ->condition('aie.changed', $last_changed, '>')